Upload
donga
View
237
Download
0
Embed Size (px)
Citation preview
NUMERICAL SIMULATION OF THE PUNCHING SHEAR BEHAVIOUR OF SELF-
COMPACTING FIBRE REINFORCED FLAT SLABS
Marcos D.E. Teixeira a, Joaquim A.O. Barros b, Vítor M.C.F. Cunha c, Bernardo N. Moraes-Neto d, António Ventura-
Gouveia e
a University of Minho, ISISE, Dep. Civil Eng., School of Eng., Campus Azurém, Guimarães, Portugal,
b University of Minho, ISISE, Dep. Civil Eng., School of Eng., Campus Azurém, Guimarães, Portugal,
c University of Minho, ISISE, Dep. Civil Eng., School of Eng., Campus Azurém, Guimarães, Portugal,
d University of Minho/University of Brasília-UnB, Dep. Civil Eng., School of Eng., Campus Azurém, Guimarães,
Portugal, [email protected]
e Instituto Politécnico de Viseu, ISISE, Dep. Civil Eng., School of Technology and Management of Viseu, Campus
Politécnico, Viseu, Portugal, [email protected]
Abstract
This paper presents the numerical simulations of the punching behaviour of centrally loaded steel fibre reinforced self-
compacting concrete (SFRSCC) flat slabs. Eight half scaled slabs reinforced with different content of hooked-end steel
fibres (0, 60, 75 and 90 kg/m3) and concrete strengths of 50 and 70 MPa were tested and numerically modelled. Moreover,
a total of 54 three-point bending tests were carried out to assess the post-cracking flexural tensile strength. All the slabs
had a relatively high conventional flexural reinforcement in order to promote the occurrence of punching failure mode.
Neither of the slabs had any type of specific shear reinforcement rather than the contribution of the steel fibres. The
numerical simulations were performed according to the Reissner-Mindlin theory under the finite element method
framework. Regarding the classic formulation of the Reissner-Mindlin theory, in order to simulate the progressive damage
induced by cracking, the shell element is discretized into layers, being assumed a plane stress state in each layer. The
numerical results are, then, compared with the experimental ones and it is possible to notice that they accurately predict
the experimental force-deflection relationship. The type of failure observed experimentally was also predicted in the
numerical simulations.
Keywords: Flat slab; Punching; Steel fibre reinforced self-compacting concrete; FEM analysis.
1. INTRODUCTION
The use of reinforced concrete (RC) flat slabs supported directly on RC columns is a well-accepted structural system
within the building construction, since it has a quite reasonable construction flexibility allowing spans with moderate
lengths between columns of about 4.5 to 6.0 m [1, 2]. Considering their constructive and architectural advantages, flat
slabs are mostly used in office, residential and parking buildings [3]. The flexural resistance of this structural system can
be achieved with relatively small thicknesses for the slab’s height. However, in order to avoid the brittle punching shear
failure mode some specific design measures are required. The latter could be overcome with either the introduction of
capitals or the inclusion of shear reinforcement. Nevertheless, these solutions are a step down on the construction
efficiency. The design of these slabs is governed by both serviceability conditions and ultimate limit states under punching
shear [4]. According to the available research, the punching behaviour of RC flat slabs can be improved by adopting
distinct strategies such as: using a high strength class concrete [5], additional shear reinforcement [6] and discrete fibres
(in particular steel fibres) [7-9]. When a hybrid reinforcement is used (i.e. rebar + discrete fibres), the steel fibres mainly
confer the punching resistance, whilst the steel bars guarantee the flexural reinforcement [10].
Fibre reinforced concrete (FRC) is mostly applied in industrial pavements, precast elements and in tunnelling. When used
in large structural elements, FRC can contribute to cost reductions, since the man-labour needed for placing the
conventional reinforcement is decreased (due to the fact that the reinforcement density is smaller when compared with
conventional RC) [11] or even eliminated if steel fibres replace all the conventional reinforcement. Better crack resistance,
ductility and toughness are among the improvements provided by fibre reinforcement. It is widely acknowledged that
there are several advantages in using FRC rather than plain concrete, i.e. greater residual tensile strengths, resistance to
fatigue, impact, blast loading and abrasion. Moreover, when comparing FRC with conventional mesh / bar reinforced
concrete, there are also some benefits such as: increased ultimate flexural and shear strength, increased toughness fatigue
and impact resistance, and better stress redistribution in the concrete composite. The application of FRC for retrofitting
and seismic design purposes has also a particular interest [12-14]. For slabs on-grade, the exclusive use of steel fibre as
reinforcement could be a viable and economical alternative to conventional reinforced concrete. According to [15, 16]
fibre contents ranging from 20 to 60 kg/m3 should be used for this structural system. On the other hand, for suspended
slabs, the exclusive use of steel fibre as reinforcement is still scarce. The outcome of FRC with enhanced mechanical
strengths has been pushing the span limits upwards. Recent pilot experiences [17-19] show the feasibility of building
elevated FRC slabs with 5-8m spans (using high percentages of fibres). The results obtained so far in these pilot
experiences appear promising, but they fail to address several important issues.
If fibres are properly selected in terms of type and content, they can provide effective reinforcement mechanisms to avoid
the punching brittle collapse [20]. The fibre reinforcement effectiveness is not solely dependent of the type and content
of fibres, the casting procedure and both rheological and mechanical properties of the concrete also influence the
reinforcement effectiveness. Since steel fibres have the highest specific weight amongst the steel fibre reinforced concrete
(SFRC) constituents, these fibres have a tendency to segregate towards the bottom surface, leading to a lower fibre content
near the top surface [21, 22]. In concrete matrices with high flowability, such is the case of self-compacting concrete
(SCC), the effect of fibre segregation is mainly due to the rheological properties of the matrix in the fresh state [23]. The
fibre segregation may lead to an anisotropic behaviour with different flexural capacity for positive and negative
curvatures.
Several authors have proposed analytical expressions to estimate the contribution of fibres for the punching resistance
based on the results obtained in experimental programs. Harajli et al. [24] have suggested a design equation to predict the
increase in ultimate punching capacity of concrete slabs due to the addition of deformed steel fibres. This simple equation
is based in the experimental results obtained by these authors and another ones available in the literature. A theoretical
approach to model the punching resistance of interior slab-column connections using SFRC is presented by Choi et al
[25]. According to Nguyen-Minh et al. [9] this formulation does not assure a good agreement with their experimental
results. More recently, Ragab [26] has simulated numerically nine steel fibre reinforced self-compacting concrete
(SFRSCC) flat slabs by performing material nonlinear analysis with software based on the finite element method (FEM).
Moreover, this author carried out a parametric study to analyse the influence of the ratio of flexural reinforcement (for
both positive and negative bending moments) and the volume of steel fibre in the slab’s behaviour. In this model, the steel
fibre reinforced concrete matrix is simulated by solid elements, while the steel bars were simulated by a link element. It
is possible to notice that this model has predicted with good accuracy the experimental results.
In this work, punching shear test results and numerical simulations are presented and discussed. Eight half scaled slabs
reinforced with different content of hooked-end steel fibres (0, 60, 75 and 90 kg/m3) and concrete strengths (50 and 70
MPa) were tested and numerically modelled. Moreover, a total of 54 three-point bending tests were carried out to assess
the post-cracking flexural tensile strength. All SFRC slabs were heavily reinforced with conventional flexural
reinforcement in order to promote the occurrence of punching failure mode. Neither of the slabs had any type of specific
shear reinforcement rather than the contribution of the steel fibres. The numerical simulations were performed by adopting
a FEM approach based on the Reissner-Mindlin shell theory. Regarding the classic formulation of the Reissner-Mindlin
shell theory, to simulate the progressive damage induced by cracking propagation, the shell element is discretized into
layers, being assumed a plane stress state in each layer. For the two out-of-plane shear components a total approach is
adopted, whereas for the in-plane components it was adopted an incremental approach. The numerical results are
compared to the experimental ones, and good predictions in terms of force-deflection relationship were obtained. The
type of failure observed experimentally was also predicted in the numerical simulations.
2. EXPERIMENTAL PROGRAM
2.1 Concrete properties, specimen geometry and test setup
Eight half scaled flat slabs prototypes with the dimensions of 2550 × 2550 × 150 mm3 (Figure 1) were centrally and
vertically loaded. Two distinct ratios for the conventional reinforcement were adopted, one for the slabs’ series with plain
concrete and another for the steel fibre reinforced self-compacting concrete (SFRSCC) slabs’ series, but the reinforcement
ratio in the critical punching region was the same in both configurations (ρ = 0.88%), Figure 2.
Eight different concrete compositions were grouped in two strength classes with compressive strengths of 50 and 70 MPa,
in order to assess its influence in the punching resistance. For similar purposes, four distinct fibre contents, 0, 60, 75 and
90 kg/m3 were adopted. The selected hooked ends steel fibres had the following mechanical and geometric characteristics:
37 mm of length (lf), diameter of 0.55 mm (df), 67 of aspect ratio (lf/df), and tensile strength of about 1100 MPa.
Hereinafter, to easily differ the distinct slabs, the acronym CfXfcY will be used, where X is the theoretical content of fibres
(per concrete cubic meter) and Y the target concrete compressive strength (in MPa). For example, the slab Cf60fc70
indicates a concrete slab planned to have a fibre content of 60 kg/m3 and a compressive strength of 70 MPa. Additionally,
for each fibre reinforced slab, nine prismatic specimens with 150×150×600 mm3 were cast and cured in the same
conditions of the corresponding slab in order to assess the residual flexural tensile strengths. The latter were assessed by
three-point bending tests carried out according the recommendation of EN 14651:2005 [27].
Table 1 presents the composition of the concrete batches. There are some differences in the compositions, since it was
necessary to have into consideration the interference of the fibres in the skeleton arrangement of the aggregates, without
compromising the requisites of self-compactability and occurrence of segregation for the constituents in all mixes. The
mix design has followed the strategy adopted by Barros et al. [28]. The mechanical properties of the steel bars were
determined by executing experimental tests according the recommendation of ISO 15630-1 [29]. Table 2 presents the
average results of the uniaxial tensile tests performed on five bar specimens.
Figure 3 illustrates the test setup, where the supports location and the load actuator position can be seen. The load actuator
had a bearing capacity of 1000 kN. The punching test was carried out under a closed-loop displacement control with a
displacement rate of 0.05 mm/min measured at the centre of the slab. The supports consisted in twelve Dywidag® steel
bars of 35 mm diameter. These bars passed through the holes of the testing slab and the holes of a stiff reaction floor. On
the top face of the holes executed in the testing slab, plates of 200 × 200 × 50 mm3 were used to anchor the extremities
of the Dywidag® bars.
Before discussing the experimental results, previously, there are some considerations that should be addressed for a better
comprehension of these results, as well as the mechanical properties for the tested concrete slabs to be adopted in the
numerical simulations. The concrete compositions and casting of the slabs were executed in a ready mix-plant, in which
the introduction of concrete constituents in the mixer was fully automated. Along with other constituents, the fibres were
directed towards the mixer in a treadmill. However, in some slabs fibres fell down from the treadmill during the path up
to the mixer, resulting in a content of fibres smaller than the expected. Therefore, in order to verify the real fibre content
(Cf), a strategy similar to the one used by Barros and Antunes [21] was adopted. This strategy consists in obtaining the
number of fibres at the fracture surface (Nf) of the three-point notched beam bending tests. Having into account the Nf
obtained from the three-point bending tests, and adopting the fibre orientation factors proposed by Abrishambaf et al.
[30] for SFRSCC with similar rheological properties and type of fibres, the Cf for the tested slabs could be inferred
according to the expression proposed by Krenchel [31]. Table 3 includes the fibre content in the tested slabs according to
this methodology. In general, the real fibre content was lower than the expected one. The Cf60fc50 and Cf75fc50 series
seem to have a significantly lower fibre content than the expected value, respectively, 50% and 57%. On the other hand,
in the Cf75fc70 composition the determined fibre content is nearby 15% higher than the target value. These occurrences
will help to explain and corroborate further ahead some discrepancies on the slab’s behaviour in the punching tests, as
well as justify some of the observed failure mechanisms.
2.2 Experimental results
Figures 4a and 4b show the load – deflection relationships, P , for the series with a compressive strength of 50 and
70 MPa, respectively. The deflection was assessed at the centre of the slabs. Table 4 resumes the relevant results
withdrawn from the P curves, where maxP is the maximum applied load and maxP is its corresponding deflection.
max
SFRSCCP is the maximum load of a SFRSCC slab and REFPmax of its correspondent reference slab; max
SFRSCC and REF
max are
the deflections at maxP of a SFRSCC slab and its correspondent reference slab. Moreover, this table also includes the
failure mode observed for each slab. In general, with the increase of the fibre content an increase of both maxP and maxP
was observed. Thus the addition of fibres contributed to an increase of the bearing capacity of the current structural
system, as well as to the increase of the deflection at maxP , i.e. toughness. The fibres proven to be more efficient on
increasing the toughness rather than the loading bearing capacity. This was due mainly by avoiding the punching failure
mechanism, as fibres contributed to the increase of the shear strength (as only specific reinforcement), while the
conventional rebars provided the main flexural reinforcement.
Due to a lower content of fibres than the expected, in the Cf60fc50 and Cf75fc50 slabs (around 30 kg/m3, see Table 3), they
have exhibited a punching failure mode, likewise the correspondent reference slab, Table 4. Notice that the reference
slabs only had flexural reinforcement and for that reason it was expected to fail by punching. The results in Table 4 show
that for the slabs with a compressive strength of 50 MPa, the increase in terms of maximum load was only significant
(12%) in the Cf90fc50, which has failed in bending. Meanwhile, for the series with a compressive strength of 70 MPa, all
the slabs reinforced with fibres have shown a higher increase in terms of maximum load, ranging from 16 to 24%.
Therefore, from the analysis of the results can be pointed out that up to a fibre content of 30 kg/m3, there is not an increase
of the slabs’ bearing capacity and is not possible to avoid the brittle punching shear failure. In terms of deflection at
maximum load, maxP , the increase of the
max max/SFRSCC REF ratio, as expected, was only significant in the slabs failing in
bending, with an increase of 72% in Cf90fc50 slab, and more than 100% in Cf90fc70. The unexpected low content of fibres
present in the Cf60fc50 and Cf75fc50 slabs (about 30 kg/m3) has only contributed to a marginal increase of the deflection
at punching failure load, nearby 4 to 6%. Thus only for fibre contents higher than 30 kg/m3 the deflection 𝛿𝑚𝑎𝑥𝑆𝐹𝑅𝑆𝐶𝐶 was
significant.
The post-cracking residual flexural tensile strength parameters (fR,i) of the distinct SFRSCC were determined with
prismatic specimens according the recommendations of EN 14651:2005 [27]. From each three-point notched beam
bending test (3PBT) it was obtained a relationship between the applied force and the crack mouth opening displacement
(CMOD). By evaluating the force at CMOD values of 0.5, 1.5, 2.5 and 3.5 mm, the fR,i values were determined by applying
the following equation:
22
3
sp
ii,R
hb
LFf
(1)
where b (=150 mm) and L (=500 mm) are the width and the span of the specimen, respectively. hsp is the distance between
the tip of the notch and the top of the cross section. In an attempt of assuring an unique crack and avoiding coalescence
of multiple macro-cracks along the notched plane in the three point SFRSCC beam bending tests, the depth of the notch
was increased with the content of fibres (25, 42.5 and 60 mm for the Cf=60, 75 and 90 kg/m3, respectively), as indicated
in Table 5. Due to technical issues only seven specimens were successfully tested in both Cf75fc70 and Cf90fc70 series.
Table 5 includes the average values of fR,i, the values in round brackets correspond to the computed coefficient of variation.
In general, relatively high values of fR,i were obtained, the two exceptions were the Cf60fc50 and Cf75fc50 series, which
was in accordance with the observed low fibre content in those series, as previously discussed. Moreover, in spite of the
high coefficients of variation observed, it was within the expectable range for this kind of material.
Figure 5 plots the relationships between the maximum load of each slab, Pmax, and the fR,i. Additionally, linear regression
curves between Pmax, and fR,i for each class strength series were also included. Good correlation factors between Pmax, and
the fR,i were found, evidencing a linear trend between the residual flexural parameters and the maximum load of the tested
slabs. This also justifies the punching failure mechanism observed in some steel fibre reinforced slabs, i.e. Cf60fc50 and
Cf75fc50, since these series unveiled the lower residual flexural strengths, which ranged between 1.96 and 5.92 MPa, and
from 2.16 to 6.35 MPa, respectively.
3. NUMERICAL SIMULATIONS OF THE PUNCHING TESTS
3.1 Numerical model
The Reissner-Mindlin theory was selected to simulate the material nonlinear behaviour of thin FRC panels, where out-
of-plane shear deformation might be appreciable [32]. The damage due to crack formation and propagation was simulated
by discretizing the thickness of the shell into several layers. Fibres bridging micro-cracks contribute to the formation of
diffuse crack patterns, since they offer resistance to the coalescence of those micro-cracks into macro-cracks. Therefore,
smeared crack constitutive models are, conceptually, more appropriate than discrete crack models in the simulation of
crack propagation in FRC structures, mainly, in those with a high number of redundant supports [33]. In the present
section only the innovative aspects implemented in the multi-fixed smeared crack model for the modelling of the punching
shear behaviour will be described. This model is implemented in the FEMIX 4.0 software, and its full description can be
found elsewhere [33]. For the case of cracked concrete the constitutive law is defined by the following equation:
0
0
crco
mf mfmf
crco
s ss
D
D
(2)
where mf and
mf are the vectors corresponding to the in-plane stress and strain increment components, respectively
(membrane and bending components); whereas s and
s are the vectors corresponding to the out-of-plane shear stress
and shear strain components, respectively. In Equation (2), crco
mfD is the in-plane cracked concrete constitutive matrix [34].
Figure 6 defines the fracture mode I modulus, cr
ID , where i and
i are the parameters that define the shape of the
crack’s normal stress vs. normal strain diagram. The ultimate crack strain ( ,
cr
n u ) is defined as a function of i and
i
parameters, fracture energy (I
fG ), tensile strength (ctf =
cr
n,1 ) and crack bandwidth (bl ) [34]. Regarding fracture mode
II, in-plane shear mode, a decrease of the shear retention factor, β, with the increase of the crack strain was adopted [34,
35]. On the other hand, Figure 7 depicts the diagram used to simulate the out-of-plane shear softening behaviour.
3.2 Two-dimensional layered geometry model
To simulate the punching tests, only one quarter of the slab was modelled, due to its double symmetry. Figure 8 depicts
both the supports adopted on the experimental setup as well as the ones used to ensure the symmetry conditions, which
are represented by blue markers. The three supports outside the symmetry axis (represented by red markers in Figure 8)
have only one fixed degree of freedom, in the zz direction, i.e. perpendicular to the xy plane that corresponds to the middle
surface of the slab. On the other hand, the blue markers located on two edges of the quarter’s slab correspond to the
supports introduced to enforce the double symmetry conditions. The load was applied at the bottom left corner of the
mesh, i.e. centre point of the slab, under displacement control by using the arc-length method.
The adopted mesh is comprised by 196 Reissner-Mindlin 8-noded shell elements (Figure 8a). A Gauss-Legendre
integration scheme of 2×2 was adopted. Each shell element was discretized into fifteen layers with a total thickness of
150 mm. Within these layers, one is used to model the steel bars non-linear behaviour, whereas the other ones are used to
model the fracture process of the SFRSCC, namely, four for the cover concrete on the top of the slabs and ten for the
remaining concrete under the steel bars. Moreover, two simultaneous smeared cracks can arise at each integration point.
In Figure 2 it is possible to notice that there are three zones of distinct longitudinal reinforcement ratios. These three zones
were taken into account in the numerical model by adopting a distinct layer discretization for each zone (Figure 8a). Thus,
Zone 1 is the one with the highest level of reinforcement, at the centre of the slab; Zone 2 is the one of intermediate level
of reinforcement; and the Zone 3 presents the lower level of reinforcement. Note that in Zone 3, for the SFRSCC slabs
there is no conventional reinforcement, so, in that zone, there are 15 layers with the same thickness, 10 mm. Table 6
includes both the thickness and corresponding number of layers in each shell element at the distinct reinforcement zones.
The layers are enumerated in the positive direction of z axis, i.e. from bottom to top slab surface.
3.3 Parameters for the constitutive material models
The mechanical properties of the steel rebars were directly derived from tensile tests, which were previously presented.
The tensile behaviour of the steel was modelled with a quadrilinear relationship, similar to the one presented in Figure 6.
However, the post-peak parameters translate the steel behaviour in the yielding stage. The parameters for the quadrilinear
diagram are presented in Table 7. Note that I
fG is a fictitious parameter that represents the energy absorption capacity of
the steel. Moreover, for the Poisson’s ratio (ν) was adopted a null value, 209.49 GPa for the Young modulus ( SE ) and
467.01 MPa for both the compressive ( cf ) and tensile yielding ( syf ) stresses.
Table 8 presents the values of parameters used to define the concrete’s constitutive post-cracking law depicted in Figure
6. For the elastic properties, a Poisson’s ratio of 0.15 and a Young modulus of 37.20 GPa were adopted. The post-cracking
behaviour considered for the Cf0fc50 and Cf0fc70 series, i.e. plain concrete, was obtained according to the
recommendations proposed by the CEB-FIP Model Code 2010 [36]. On the other hand, the parameter values used for the
SFRSCC post-cracking behaviour were obtained by inverse analysis by fitting the numerical response to the experimental
curves obtained in the punching tests. Details of this procedure can be found elsewhere [37].
It is also important to refer that for both slabs Cf60fc50 and Cf75fc50, the concrete layers were ascribed with distinct tensile
post-cracking behaviour as resumed in Table 8. As previously detailed, these slabs had some issues during their casting,
in particular with the effective fibre content introduced into the mixture. Moreover, to take into account a heterogeneous
fibre distribution along the slab’s depth, which could be ascribed to fibre segregation problems, distinct values for the
fracture energy were attributed to the layers. Nevertheless, the parameters i and i were kept constant, therefore the
shape of the post-cracking diagram was similar within the slab’s distinct layers. Three cylindrical cores were performed
in each Cf60fc50 and Cf75fc50 slabs in order to assess the variation of fibre content along the slab’s height. The fracture
energy variation along the slab’s height was defined according to the fibre content variation. The fibre content followed
up a parabolic trend from the bottom to the top layer. A fracture energy decrease from 1.25 to 0.08 N/mm with the same
trend of the fibre content was ascribed from the bottom, layer n. 1 (compressed zone), to the top layer, layer n. 15
(tensioned zone), respectively (Figure 8b and Table 8). Since the real fibre content and average compressive strength of
Cf75fc50 slab were similar to those of Cf60fc50 slab, the same values for the constitutive post-cracking relationship were
adopted in the simulation of these two slabs. Notice also that the value of the fracture energy for the slab Cf75fc70 (with
an expected fibre content of 75 kg/m3) is higher than the fracture energy for the slab Cf90fc70 (with an expected fibre
content of 90 kg/m3). This could be explained by the real fibre content present in each slab, which was lower in the slab
Cf90fc70, see Table 3.
3.4 Comparison of the numerical and experimental results
Figures 9 and 10 depict, for each slab, both the numerical and experimental curves. The numerical simulations are in good
accordance with the experimental curves. In general, the observed failure modes on the experimental tests were
successfully predicted. Concerning the numerical simulation of the Cf0fc50 slab (Figure 9a), it was predicted the punching
shear failure, however the displacement for which this type of failure occurs was slightly overestimated, approximately
10%. On the other hand, for the Cf60fc50 and Cf75fc50 slabs (Figure 9b) the load decay observed experimentally due to
punching was not captured numerically due to difficulties on the convergence process of the numerical simulations of
both slabs. However, the deformational response was predicted with high accuracy, and since convergence was not
possible to assure precisely at the deflection level coinciding with the punching failure observed experimentally, it is
presumed that the model is almost in the situation of capturing this type of failure mode. Finally, in the Cf90fc50 slab
(Figure 9c), both the experimental and numerical curves are almost coincident in all their development. Neither of them
show a punching failure, due to the high fibre content used in this series.
On the other hand, for the series with a compressive strength of 70 MPa, the simulations were even more accurate. As it
can be seen in Figure 10, all the simulations were almost coincident with the experimental curves. Once again, for the
slab Cf0fc70 (Figure 10a) the punching failure was not precisely captured, again due to convergence problems, however,
the last converged point of the numerical curve is practically coincident with the experimental punching failure. For both
Cf60fc70 (Figure 10b) and Cf75fc70 (Figure 10c), the experimental curves do not exhibited a punching failure and neither
did the numerical ones. Moreover, they were almost coincident on all their extent. The numerical curve regarding the slab
Cf90fc70 (Figure 10d) showed that the experimental and numerical curves were almost coincident and the
bending/punching failure was numerically predicted for the same displacement and level of loading.
4. PARAMETRIC ANALYSIS
As previously stated, due to some issues that occurred during the casting procedure, some slabs exhibited a lower fibre
content than it was expected. Therefore, it could not be fully ascertained for what level of fibre content and concrete
strength the punching shear failure could be avoided. Moreover, the length of the selected fibre was relatively short
(37 mm) and its aspect ratio not so high (67), thus the selected fibre may not be the most effective one. It is well established
that a higher aspect ratio will lead to a higher number of fibres in concrete bulk [37], which increases the probability of
having higher number of fibres bridging active cracks. Hence, in this section will be conducted a parametric study to
assess the influence of the fracture energy parameter and the shape of the second branch of the post-cracking stress –
strain diagram on the deformational response of the slabs. The fracture energy will be obtained according to experimental
data for fibres with a higher aspect ratio [33]. Having in mind that the numerical model was previously calibrated and
shown a good accuracy on predicting the experimental results, the outcome of the numerical simulations comprised in
the parametric study may render a relatively accurate estimation of the behaviour of slabs reinforced with fibres of a
higher aspect ratio and distinct fibre contents.
4.1 Influence of the fracture energy
The influence of the fracture energy on the deformational response of the slab was ascertained. For this purpose, the
fracture energy of concretes corresponding to the fibre contents, Cf, adopted in the experimental programme, i.e. 60, 75
and 90 kg/m3, were assessed through an analytical expression proposed by Barros and Figueiras [33]. The latter correlates
the fracture energy of a fibre reinforced concrete with the weight percentage of fibres (60/.80). This fibre has a length of
60 mm, diameter of 0.8 mm and an aspect ratio of 75. Thus, the fracture energy of fibre reinforced concrete, I
fG , was
computed as follows:
827.1159.130.1
0
fW
fG
I
fG
(3)
where 0fG is the fracture energy of the corresponding plain concrete and fW is the weight percentage of fibres in the
mixture. According to the CEB-FIB Model Code 2010 [36], for a plain concrete with an average compressive strength of
50 MPa, the value of fracture energy is 0.147 N/mm, whilst for a plain concrete with a compressive strength of 70 MPa,
the fracture energy is 0.156 N/mm. Table 9 presents the estimated fracture energy for fibre reinforced concrete
compositions with three distinct fibre contents, i.e. 60, 75 and 90 kg/m3, and using a fibre with a higher aspect ratio (75).
Hence, the numerical simulations were carried out using the parameters of the post-cracking stress – strain diagram used
for the Cf60fc50 and Cf60fc70 series, changing only the fracture energy value by the ones computed with Equation (3) and
presented in Table 9. Figures 11a and 11b show the numerical simulations carried out for the series with a compressive
strength of 50 and 70 MPa, respectively, assuming the new estimated fracture energy values presented in Table 9
computed for the new hooked end steel fibres (60/.80). As it was expected, the load carrying capacity and toughness of
the slabs increased with the increase of the fracture energy, regardless of the compressive strength adopted in this study.
Table 10 presents the ratio between the maximum numerical load ( maxP ) and the maximum experimental load (ExpPmax ),
the ratio between the numerical displacement at maximum load ( max ) and the correspondent experimental one (Exp
max ),
and the ratio between the numerical load for a crack width of 0.3 mm ( mmP 3.0 ) and the correspondent experimental one
(Exp
mmP 3.0 ). These values were obtained considering the use of 60, 75 and 90 kg/m3 of fibres 60/.80. Note that the load
mmP 3.0 was obtained for the numerical load step in which a crack reached the 0.3 mm width. In the present case this
occurred for the integration point nearer the load application and in the top layer of the Mindlin shell element. Moreover,
since there was no experimental information regarding the crack opening, for the series Exp – Cf60, it was used the
information from the numerical curve obtained by inverse analysis in section 4.2. The crack width was obtained by
multiplying the crack normal strain component by the crack band width.
In what concerns to numerical results assuming a fibre content of 60 kg/m3 with a fibre length of 60 mm and an aspect
ratio of 75, when compared to the experimental results (60 kg/m3 with a fibre length of 37 mm and an aspect ratio of 67),
it is visible an increase of both the maximum load and load at the onset of a crack opening of 0.3 mm. This increase was
more significant for the series with a compressive strength of 70 MPa, which was about 26%. Thus, the use of a fibre
with a higher aspect ratio would assure higher load carrying capacity for this flat slabs structural system at both
serviceability and ultimate stages. It was also noticed for a fibre content of 90 kg/m3 an increase in the maximum load
about 86 and 69%, respectively, for the series with 50 and 70 MPa. The increase of the displacement at the maximum
load was also significant for the series Cf90, and was near 79 and 47%, respectively, for the series with 50 and 70 MPa.
4.2 Influence of the shape of the material law
The influence of the slope of second branch of the stress – strain diagram on the response of the slabs was assessed. For
this purpose, it is considered the same fracture energy for all the simulations carried out, but for the stress of the second
post-peak point for the following three distinct levels of 0.50fctm, 1.00fctm and 1.25fctm were considered, where fctm was
the average concrete tensile strength. The post-cracked stress-strain diagrams used in the numerical simulations are
depicted in Figures 12a and 12b for the 50 and 70 MPa series, respectively.
The values of the fracture energy considered correspond to the ones computed for the slabs Cf60fc50 and Cf60fc70 (Table
7). Figures 13a and 13b depict the results of the simulations. The load carrying capacity of the slabs increased with the
increase of the stress value of the point defining the shape of the second branch of the trilinear diagram. Table 11 includes
the max max
ExpP P , max max
Exp and 0.3 0.3
Exp
mm mmP P ratios obtained for the three stress values adopted for the second point
of the trilinear diagram, respectively, for the fcm=50 and 70 MPa series. It can be observed that the maximum load was
higher than the one observed in the experimental results using the fibres with a lower aspect ratio. For the 50 MPa series
the increase of the load was in-between 8% and 12%, while for the 70 MPa series this increase ranged between 6% and
21%. Table 11 includes also the ratio between the numerical load for which was attained a crack width opening of 0.3
mm and the correspondent experimental load. For the same values of Gf and by modifying the inclination of the second
branch, the load at a crack opening width of 0.3 mm was 16 to 31% higher than the correspondent one for the experimental
curve, for the series with a compressive strength of 50 MPa. On the other hand, for the higher strength concrete (70 MPa)
a smaller increase ranging 2% and 18% was observed for the load at a crack opening width of 0.3 mm.
It is also important to refer that the influence of shifting horizontally the second post-peak point of the stress-strain
diagram, i.e. changing the value of the ξ2 parameter (Figure 6), was also studied. Considering the new horizontal position
as 1.40 ξ2 and 2.00 ξ2 (for the 50 MPa series), and 0.875 ξ2 and 1.125 ξ2 (for the 70 MPa series) of the initial ξ2 parameter
no significant differences on the slab’s response were observed, since all the numerical curves were almost coincident.
This could be explained by the fact that for the attained slab’s deflection level, the strain at the integration point did not
reach the strain corresponding to the second post-peak point of the stress-strain diagram. Moreover, this could also suggest
that the slab’s flexural residual capacity is still far from being reached.
5. CONCLUSIONS
In this work the results of punching shear tests carried out on flat slabs with a hybrid reinforcement (rebar + discrete
fibres) are presented and discussed. Eight half scaled slabs, reinforced with distinct fibre contents, namely, 0, 60, 75 and
90 kg/m3, and different compressive strengths (50 and 70 MPa) were tested. The experimental results were numerically
modelled under the FEM framework. Moreover, a total of 54 three-point bending tests were carried out to assess the post-
cracking flexural tensile strength.
In the experimental tests, and due to deficiencies on the automatic procedure of introducing the fibres into the mixer, the
slabs reinforced with 60 and 75 kg/m3 and a compressive strength of 50 MPa showed relative low content of fibres (around
30 kg/m3). Therefore, both of them have failed in punching like the corresponding reference slabs (without fibre
reinforcement). For the slabs reinforced with the highest fibre content (i.e 90 kg/m3) an increase from 12 to 24% in terms
of maximum load was observed, followed by a flexural failure mode. The increase in the maximum deflection, maxP ,
was only significant in the slabs with the highest fibre content, with an increase between 72 and 102%.
The load-deflection curves obtained in the numerical simulations are in good accordance with the experimental ones. In
general, the type of failure mode observed on the experimental tests was also successfully predicted. A parametric study
was also carried out in order to assess the influence on the slab’s mechanical response of the fracture energy and the shape
of the first branch of the diagram that defines the crack opening process. The values of the fracture energy were estimated
having in consideration the use of a fibre with a higher aspect ratio. It was found out that the higher aspect ratio lead to
higher fracture energies, and consequently, the slab presented a higher load carrying capacity without occurring the
punching shear failure. In terms of the shape of the first branch of the crack constitutive law, it was observed that when
the first post-cracking stress point is 1.25fctm (tensile strain hardening), a significant increase occurs on the load carrying
capacity for both serviceability and ultimate limit states.
6. ACKNOWLEDGEMENTS
This work is supported by FEDER funds through the Operational Program for Competitiveness Factors - COMPETE and
National Funds through FCT - Portuguese Foundation for Science and Technology under the project “SlabSys – HFRC –
Flat slabs for multi-storey buildings using hybrid reinforced self-compacting concrete: an innovative structural system”
PTDC/ECM/120394/2010.
7. REFERENCES
[1] Wight, J.K.; MacGeregor, J.G. Reinforced Concrete: Mechanics and Design. Fifth Edition, New Jersey: Prentice Hall,
2009.
[2] Delahay, P.E.J.; Christopher, P.E.B. Current trends in economical concrete construction, Part one: Floor framing and
lateral systems. Structure Magazine, pp. 19-21, July 2007.
[3] Maya, L. F.; Fernández Ruiz, M.; Muttoni, A.; Foster, S. J. Punching shear strength of steel fibre reinforced concrete
slabs. Engineering Structures, 40(0), pp. 83-94, 2012.
[4] Muttoni A. Punching Shear Strength of Reinforced Concrete Slabs without Transverse Reinforcement. ACI Structural
Journal, 105(4), pp. 440-450, 2008.
[5] Ozden, S.; Ersoy, U.; Ozturan, T., Punching shear tests of normal- and high-strength concrete flat plates. Canadian
Journal of Civil Engineering, 2006. 33(11): p. 1389-1400.
[6] Melo, G.S.; Coelho, A.E.G.; Oliveira, D.R.C. Reinforced concrete flat slabs with inclined stirrups as shear
reinforcement. International Workshop on Punching Shear Capacity of RC Slabs – Proceedings. TRITA-BKN, Bulletin
57, pp. 155-162, 2000.
[7] Tan, K-H.; Paramasivam, P. Punching shear strength of steel fiber reinforced concrete slabs. Journal of Materials in
Civil Engineering, 6(2), pp. 240-253, May 1994.
[8] Prisco, M.; Felicetti, R. Some results on punching shear in plain and fibre-reinforced micro-concrete slabs. Magazine
of Concrete Research, 49(180), pp. 201-219, September 1997.
[9] Nguyen-Minh, L.; Rovňák, M.; Tran-Quoc, T.; Nguyenkim, K. Punching Shear Resistance of Steel Fiber Reinforced
Concrete Flat Slabs. Procedia Engineering 14(0): 1830-1837, 2011.
[10] Moraes-Neto, B.N.; Barros, J.A.O.; Melo, G.S.S.A. A model for the prediction of the punching resistance of steel
fibre reinforced concrete slabs centrically loaded. Construction and Building Materials Journal, 46, pp. 211-223, 2013.
[11] Soutsos, M. N.; Le, T. T.; Lampropoulos, A. P. Flexural performance of fibre reinforced concrete made with steel
and synthetic fibres. Construction and Building Materials 36(0): 704-710, 2012.
[12] Banthia, N. in: J.I. Daniel, S.P. Shah (Eds.), ACI SP-142. Fiber Reinforced Concrete ACI, pp. 91–120, 1994.
[13] di Prisco, M.; Plizzari, G.; Vandewalle, L. Fibre reinforced concrete: new design perspectives. Materials and
Structures 42(9): 1261-1281, 2009.
[14] Balaguru, P.N.; Shah, S.P. Fiber Reinforced Cement Composites, McGraw Hill, 1992.
[15] Michels, J.; Waldmann, D.; Maas, S.; Zürbes, A. Steel fibers as only reinforcement for flat slab construction
Experimental investigation and design. Construction and Building Materials 26(1): 145-155, 2012.
[16] Barros, J.A.O.; Figueiras, J.A. Experimental behaviour of fiber concrete slabs on soil. Mechanics of Cohesive
Frictional Materials 3(3): 277-290, 1998.
[17] X. Destrée. Steel-fibre-only reinforced concrete in free suspended elevated slabs. Concrete Engineering International.
13(1), pp. 47-49, 2009.
[18] X. Destrée. Structural application of steel fibre as principal reinforcing: conditions – Design examples, in RILEM
PRO15, Fibre Reinforced Concretes (FRC) BEFIB 2000, 2000.
[19] X. Destrée. Steel-fibre-reinforced pile-supported slabs. Concrete (London). 41(4), pp. 26-27, 2007.
[20] Swamy, R.N. FRC for sustainable infrastructure regeneration and rehabilitation In Fibre-Reinforced Concretes (FRC)
BEFIB 2000: 3-18, Lyon, France, 2000.
[21] Barros, J.A.O.; Antunes, J.A.B. Experimental characterization of the flexural behaviour of steel fibre reinforced
concrete according to RILEM TC 162-TDF recommendations. RILEM TC 162 TDF Workshop, 77-89, March, 2003.
[22] Babut, R. Structural investigation of steel fibre reinforced concrete. Fracture mechanics and structural aspects of
concrete 31: 29-44. Delft, Netherlands, 1986.
[23] Stähli P.; Custer R.; van Mier J. On flow properties, fibre distribution, fibre orientation and flexural behaviour of
FRC. Materials and Structures, 41(1), 189-196, 2008.
[24] Harajli, M.H.; Maalouf, D.; Khatib, H. Effect of fibers on the punching shear strength of slab-column connections.
Cement and Concrete Composites 17(2): 161-170, 1995.
[25] Choi, K.K.; Reda Taha, M.M.; Park, H.G.; Maji, A.K. Punching shear strength of interior concrete slab–column
connections reinforced with steel fibers. Cement and Concrete Composites 29(5): 409-420, 2007.
[26] Ragab, K.S. Study Punching Shear of Steel Fiber Reinforced Self Compacting Concrete Slabs by Nonlinear Analysis.
International Journal of Civil Science and Engineering. 7: 106-118, 2013.
[27] EN 14651:2005. Test method for metallic fibre concrete – Measuring the flexural tensile strength (limit of
proportionality (LOP), residual). CEN - European Committee for Standardization, 2007.
[28] Barros, J.A.O.; Pereira, E.B.; Santos, S.P.F. Lightweight panels of steel fiber reinforced self-compacting concrete.
ASCE Materials in Civil Engineering Journal, 19(4), pp. 295-304, 2007. http://hdl.handle.net/1822/9246
[29] ISO 15630-1. Steel for the reinforcement and prestressing of concrete - Test methods - Part 1: Reinforcing bars, wire
rod and wire. International Standard, First edition, April 2002.
[30] Abrishambaf, A.; Barros, J.A.O.; Cunha, V.M.C.F. Relation between fibre distribution and post-cracking behaviour
in steel fibre reinforced self-compacting concrete panels, Cement and Concrete Research 51, pp. 57–66, 2013.
[31] Krenchel H. (1975). Fibre spacing and specific fibre surface. In RILEM Symposium: Fibre Reinforced Cement and
Concrete, Volume 1, pp. 69-79. Horneby, Lencaster.
[32] Reissner, E., The effect of transverse shear deformation on the bending elastic plates. Journal of Applied Mechanics,
12, pp. 69-76, 1945.
[33] Barros, J.A.O.; Figueiras, J.A. Nonlinear analysis of steel fibre reinforced concrete slabs on grade. Computers &
Structures Journal, 79(1), pp. 97-106, January 2001.
[34] Ventura-Gouveia, A.; Barros, J.A.O; Azevedo, A.F.M.; Crack constitutive model for the prediction of punching
failure modes of fiber reinforced concrete laminar structures, Computers & Concrete, 8(6), 735-755, 2011.
[35] Rots, J.G.; de Borst, R. Analysis of mixed-mode fracture in concrete. Journal of Engineering Mechanics, ASCE,
113(11), pp. 1739-1758, 1987.
[36] CEB-FIB Model Code 2010 CEB-FIB Model Code 2010 – Final draft. Thomas Thelford, Lausanne, Switzerland,
2010.
[37] Barros, J.A.O.; Cunha, V.M.C.F.; Ribeiro, A.F.; Antunes, J.A.B. Post-Cracking Behaviour of Steel Fibre Reinforced
Concrete. Materials and Structures Journal, RILEM, 38(275), pp. 47-56, 2005.
LIST OF TABLE CAPTIONS
Table 1. Self-compacting concrete mixes [per m3].
Table 2. Average values of the tensile properties of the steel bars used in the flexural reinforcement of the slabs.
Table 3. Content of fibres in the notched plane of SFRSCC beam bending tests, and the determined content of fibres.
Table 4. Main results obtained in the punching tests.
Table 5. Residual flexural strengths, fR,i, obtained in the three-point bending tests.
Table 6. Thickness of the layers in the shell element (mm).
Table 7. Parameter values used to define the steel bar quadrilinear tensile law.
Table 8. Parameter values of the post-cracking trilinear relationship for the concrete bulk.
Table 9. Estimated fracture energy of concrete compositions reinforced with 60/.80 fibres.
Table 10. Relations between the maximum load and displacement for each simulation and its experimental results (same
material law).
Table 11. Relations between the maximum load and displacement for each simulation and its experimental results (same
fracture energy).
Table 1. Self-compacting concrete mixes [per m3].
Constituent
[kg]
Series
Cf0fc50 Cf0fc70 Cf60fc50 Cf60fc70 Cf75fc50 Cf75fc70 Cf90fc50 Cf90fc70
Cement 420 480 420 480 420 480 420 480
Fly ash 65 65 65 65 75 75 80 80
Coarse aggregate
(Dmax=12 mm) 670 628 668 627 661 620 657 619
Coarse Sand 806 757 804 755 795 746 790 741
Fine sand 269 253 268 252 265 249 265 247
Superplasticizer 6.79 7.63 7.76 8.72 7.92 8.88 8.00 8.96
Water 165 185 165 185 168 189 170 190
Fibres 0 0 60 60 75 75 90 90
Table 2. Average values of the tensile properties of the steel bars used in the flexural reinforcement of the slabs.
Diameter (mm) εsy (‰) fsy (MPa) εsu (‰) fsu (MPa) Es (GPa)
12 2.53 530 - 684 209.49
Table 3. Content of fibres in the notched plane of SFRSCC beam bending tests, and the determined content of fibres.
Series Depth of the notched
cross section (mm) Nº of fibres (Nf)
Fibres per unit area
(Nf/cm2)
Determined content of fibres
(kg/m3)
Cf60fc50 125 166 0.89 30
Cf60fc70 125 295 1.58 53
Cf75fc50 107.5 155 0.96 32
Cf75fc70 107.5 416 2.58 86
Cf90fc50 90 305 2.26 76
Cf90fc70 90 317 2.34 79
Table 4. Main results obtained in the punching tests.
Slab maxP
(kN) maxP
(mm) max max
SFRSCC REFP P max max
SFRSCC REF Failure mode
Cf0fc50 409.11 26.58 - - Punching
Cf0fc70 399.15 25.09 - - Punching
Cf60fc50 386.08 28.23 0.94 1.06 Punching
Cf60fc70 462.03 38.63 1.16 1.54 Bending
Cf75fc50 408.03 27.56 1.00 1.04 Punching
Cf75fc70 496.34 48.79 1.24 1.94 Bending
Cf90fc50 456.53 45.81 1.12 1.72 Bending
Cf90fc70 495.58 50.77 1.24 2.02 Bending/Punching
Table 5. Residual flexural strengths, fR,i, obtained in the three-point bending tests.
Series N. specimens Notch depth fR,1 fR,2 fR,3 fR,4
[-] [mm] [MPa] [MPa] [MPa] [MPa]
Cf0fc50 - - - - - -
Cf0fc70 - - - - - -
Cf60fc50 9 25 5.92
(20.9%)
4.28
(19.5%)
2.74
(21.8%)
1.96
(29.9%)
Cf60fc70 9 25 9.54
(13.4%)
10.08
(10.7%)
8.37
(14.3%)
6.90
(17.5%)
Cf75fc50 9 42.5 6.35
(23.9%)
5.30
(18.1%)
3.21
(23.0%)
2.16
(31.6%)
Cf75fc70 7 42.5 12.39
(13.4%)
13.01
(11.7%)
11.74
(16.1%)
10.54
(19.3%)
Cf90fc50 9 60 12.31
(15.4%)
14.24
(16.3%)
12.77
(21.1%)
10.73
(25.3%)
Cf90fc70 7 60 11.55
(22.5%)
13.33
(21.0%)
12.08
(24.8%)
10.57
(31.0%)
Table 6. Thickness of the layers in the shell element (mm).
Slab Zone
Concrete Steel Cover Concrete
Layer Thick Layer Thick Layer Thick
Reference
1 [1-10] 12.238 11 1.244 [12-15] 6.595
2 [1-10] 12.265 11 0.693 [12-15] 6.664
3 [1-10] 12.293 11 0.141 [12-15] 6.733
SFRSCC
1 [1-10] 12.238 11 1.244 [12-15] 6.595
2 [1-10] 12.269 11 0.622 [12-15] 6.672
3 [1-15] 10.000
Table 7. Parameter values used to define the steel bar quadrilinear tensile law.
𝜎𝑛,1𝑐𝑟
(MPa) 𝜎𝑛,2𝑐𝑟
(MPa)
𝜎𝑛,3𝑐𝑟
(MPa)
𝜎𝑛,4𝑐𝑟
(MPa)
휀𝑛,2𝑐𝑟
(‰)
휀𝑛,3𝑐𝑟
(‰)
휀𝑛,4𝑐𝑟
(‰)
I
fG
(N/mm)
467.01 511.99 531.39 584.00 0.72 1.84 7.74 390.30
Table 8. Parameter values of the post-cracking trilinear relationship for the concrete bulk.
Series Layer ξ1 (-) ξ2 (-) α1 (-) α2 (-)
I
fG
(N/mm)
Cf0fc50 - 0.100 0.999 0.095 0.000 0.08
Cf60fc50 [1-10] 0.033 0.250 0.250 0.250
[1.25 – 0.16]
Cf75fc50 [12-15] 0.08
Cf90fc50 - 0.005 0.100 0.300 0.200 3.50
Cf0fc70 - 0.100 0.999 0.095 0.000 0.08
Cf60fc70 0.025 0.800 0.250 0.250 1.00
Cf75fc70 0.002 0.150 0.350 0.200 4.00
Cf90fc70 0.002 0.100 0.300 0.200 3.50
Table 9. Estimated fracture energy of concrete compositions reinforced with 60/.80 fibres.
Cf
(kg/m3)
Wf
(%)
Gf (N/mm)
fc =50 (MPa) fc = 70 (MPa)
60 2.500 10.46 11.11
75 3.125 15.66 16.62
90 3.750 21.79 23.12
Table 10. Relations between the maximum load and displacement for each simulation and its experimental results (same
material law).
Cf 60 kg/m3 75 kg/m3 90 kg/m3
fcm
(MPa) ExpP
P
max
max
Exp
max
max
Exp
mm
mm
P
P
3.0
3.0
ExpP
P
max
max
Exp
max
max
Exp
mm
mm
P
P
3.0
3.0
ExpP
P
max
max
Exp
max
max
Exp
mm
mm
P
P
3.0
3.0
50 1.163 0.826 1.154 1.526 1.207 1.177 1.857 1.793 1.311
70 1.276 0.693 1.255 1.485 1.333 1.152 1.692 1.467 1.182
Table 11. Relations between the maximum load and displacement for each simulation and its experimental results (same
fracture energy).
0.50fctm 1.00fctm 1.25fctm
fcm
(MPa) ExpP
P
max
max
Exp
max
max
Exp
mm
mm
P
P
3.0
3.0
ExpP
P
max
max
Exp
max
max
Exp
mm
mm
P
P
3.0
3.0
ExpP
P
max
max
Exp
max
max
Exp
mm
mm
P
P
3.0
3.0
50 1.120 1.207 0.974 1.088 1.000 1.164 1.089 0.828 1.315
70 1.060 1.026 1.024 1.205 1.053 1.115 1.201 0.947 1.181
LIST OF FIGURE CAPTIONS
Figure 1. Geometry of the slab prototype (dimensions in mm).
Figure 2. Flexural reinforcement in: a) reference slabs, b) SFRC slabs; Cross section of c) AA, and d) BB (dimensions in
mm).
Figure 3. Test setup: a) Top view, b) Bottom view.
Figure 4. Load versus centre deflection in the tested series: a) 50 MPa, b) 70 MPa.
Figure 5. Relationships between Pmax and the residual flexural tensile strengths: a) fR,1; b) fR,2; c) fR,3; d) fR,4
Figure 6. Trilinear stress-strain diagram to simulate the fracture mode I crack propagation.
Figure 7. Generic diagram to simulate the relationship between the out-of-plane (OP) shear stress and shear strain
components.
Figure 8. a) Mesh, supports and symmetry conditions adopted; b) Layer discretization of the slab’s cross section.
Figure 9. Experimental and numerical results (50 MPa).
Figure 10. Experimental and numerical results (70 MPa).
Figure 11. Curves load-displacement for different values of fracture energy. a) fc = 50 MPa, b) fc = 70 MPa.
Figure 12. Different material laws considered. a) fc = 50 MPa, b) fc = 70 MPa.
Figure 13. Curves load-displacement for different shapes of the material law. a) fc = 50 MPa, b) fc = 70 MPa.
Figure 1. Geometry of the slab prototype (dimensions in mm) [18].
2550
e=200
20
0
28
64
64
75
01
50
28
64
64
15
0
e=200
15
0
Edge condition
11
86
Spring support
a) b)
c)
d)
Figure 2. Flexural reinforcement in: a) reference slabs, b) SFRSCC slabs; Cross section of c) AA, and d) BB
(dimensions in mm) [18].
11Ø12@100
11Ø
12
@1
00
A A
3Ø12@240 3Ø12@240
3Ø
12
@2
40
240240
24
05
0Steel bars for
the supporting
of the reinforcement
11Ø12@100
11
Ø12@
100
B B
e+6d
50
e=200
11Ø12@100 3Ø12@2403Ø12@240 240 240
15
0
d
150
e=200
11Ø12@100
d
Steel bars for the supporting
of the reinforcement
a) b)
Figure 3. Test setup: a) Top view, b) Bottom view [18].
a) b)
Figure 4. Load versus centre deflection in the tested series: a) 50 MPa, b) 70 MPa.
0 10 20 30 40 500
100
200
300
400
500
Load
(kN
)
Displacement (mm)
Reference 60 kg/m3 (aprox. 30 kg/m
3)
75 kg/m3 (aprox. 30 kg/m
3) 90 kg/m
3
0 10 20 30 40 500
100
200
300
400
500
Load
(kN
)
Displacement (mm)
Reference 60 kg/m3
75 kg/m3 90 kg/m
3
a) b)
c) d)
Figure 5. Relationships between Pmax and the residual flexural tensile strengths: a) fR,1; b) fR,2; c) fR,3; d) fR,4
0 2 4 6 8 10 12 14 16 18 20300
350
400
450
500
550
600
P fc50
max = 9.79 f
R,1 + 336.62
R2 = 0.88
Pm
ax [
kN
]
fR,1
[MPa]
fc50-Avg fc50
fc70-Avg fc70
Linear Fit of fc50-Avg
Linear Fit of fc70-Avg
P fc70
max = 12.88 f
R,1 + 340.91
R2 = 0.86
0 2 4 6 8 10 12 14 16 18 20300
350
400
450
500
550
600
P fc50
max = 6.42 f
R,2 + 365.87
R2 = 0.91
Pm
ax [
kN
]
fR,2
[MPa]
fc50-Avg fc50
fc70-Avg fc70
Linear Fit of fc50-Avg
Linear Fit of fc70-Avg
P fc70
max = 10.90 f
R,2 + 352.37
R2 = 0.98
0 2 4 6 8 10 12 14 16 18 20300
350
400
450
500
550
600
P fc50
max = 6.15 f
R,3 + 378.51
R2 = 0.86
Pm
ax [
kN
]
fR,3
[MPa]
fc50-Avg fc50
fc70-Avg fc70
Linear Fit of fc50-Avg
Linear Fit of fc70-Avg
P fc70
max = 9.50 f
R,3 + 382.59
R2 = 0.98
0 2 4 6 8 10 12 14 16 18 20300
350
400
450
500
550
600
P fc50
max = 6.90 f
R,4 + 382.70
R2 = 0.84
Pm
ax [
kN
]
fR,4
[MPa]
fc50-Avg fc50
fc70-Avg fc70
Linear Fit of fc50-Avg
Linear Fit of fc70-Avg
P fc70
max = 9.29 f
R,4 + 397.90
R2 = 0.99
Figure 6. Trilinear stress-strain diagram to simulate the fracture mode I crack propagation ( cr
n,1 ctf , cr cr
n,2 1 n,1 , cr cr
n,3 2 n,1 , ,2 1 ,
cr cr
n n u , ,3 2 ,
cr cr
n n u ).
GI
f
lb
n,u crncr
cr
n
cr
n,1
DI,sec
cr
DI,1
cr
DI,2
cr
DI,3
cr
n,3 cr
n,2 cr
cr
n,2
cr
n,3
n,max cr
cr
n,max
Figure 7. Generic diagram to simulate the relationship between the out-of-plane (OP) shear stress and shear strain
components.
p
G
linear behavior
c DIII,sec
GIII
f
lb
OP
OP
max OP
OP
u OP
max OP
p
max- OP
- OP
u- OP
max- OP
p
p- OP
OP
a) b)
Figure 8. a) Mesh, supports and symmetry conditions adopted; b) Layer discretization of the slab’s cross section.
a) Reference Slab b) 60 kg/m3 and 75 kg/m3
c) 90 kg/m3
Figure 9. Experimental and numerical results (50 MPa).
0 10 20 30 40 500
100
200
300
400
500
Lo
ad (
kN
)
Displacement (mm)
Experimental
Numerical
0 10 20 30 40 500
100
200
300
400
500
Lo
ad (
kN
)
Displacement (mm)
Experimental 60 kg/m3
Experimental 75 kg/m3
Numerical
0 10 20 30 40 500
100
200
300
400
500
Lo
ad (
kN
)
Displacement (mm)
Experimental
Numerical
a) Reference Slab b) 60 kg/m3
c) 75 kg/m3 d) 90 kg/m3
Figure 10. Experimental and numerical results (70 MPa).
Experimental
Numerical
0 10 20 30 40 50 600
100
200
300
400
500L
oad
(kN
)
Displacement (mm)
0 10 20 30 40 50 600
100
200
300
400
500
Load
(kN
)
Displacement (mm)
Experimental
Numerical
0 10 20 30 40 50 600
100
200
300
400
500
Load
(kN
)
Displacement (mm)
Experimental
Numerical
0 10 20 30 40 50 600
100
200
300
400
500
Load
(kN
)
Displacement (mm)
Experimental
Numerical
a) fc = 50 MPa b) fc = 70 MPa
Figure 11. Load-displacement curves for distinct fracture energy values.
0 10 20 30 40 500
100
200
300
400
500
600
700
800
Load
(kN
)
Displacement (mm)
Exp - Cf60
Cf60
Cf75
Cf90
fG
Exp - Cf60
Cf60
Cf75
Cf90
0 10 20 30 40 500
100
200
300
400
500
600
700
800
Lo
ad (
kN
)
Displacement (mm)
fG
a) fc = 50 MPa b) fc = 70 MPa
Figure 12. Different material laws considered.
0.000 0.005 0.010 0.015 0.020 0.025 0.0300
1
2
3
4
5
cr n (
MP
a)
cr
n
0.50fctm
1.00fctm
1.25fctm
0.000 0.005 0.010 0.015 0.020 0.025 0.0300
1
2
3
4
5
cr n
cr
n
0.50fctm
1.00fctm
1.25fctm
a) fc = 50 MPa b) fc = 70 MPa
Figure 13. Curves load-displacement for different shapes of the material law.
Exp - Cf60
0.50fctm
1.00fctm
1.25fctm
0 5 10 15 20 25 30 35 40 45 500
100
200
300
400
500
600
Load
(kN
)
Displacement (mm)
0 5 10 15 20 25 30 35 40 45 500
100
200
300
400
500
600
Load
(kN
)
Displacement (mm)
Exp - Cf60
0.50fctm
1.00fctm
1.25fctm